面向仿真优化的粒子群算法计算模型  

Simulation Based Optimization for PSO Computational Model

作　　者：李斌[1,2] 李文锋[2] 

机构地区：[1]福建工程学院经济管理系,福州350108 [2]武汉理工大学物流工程学院,武汉430063

出　　处：《系统仿真学报》2011年第10期2118-2124,共7页Journal of System Simulation

基　　金：国家十一五科技支撑计划课题支持(2006BAH02A06);福建省教育厅A类科技项目(JA10214);教育部人文社会科学研究青年基金项目(11YJC630089)

摘　　要：粒子群优化算法(Particle Swarm Optimization,PSO)是群集智能的典型代表,其参数较少且操作简单,故一直是智能优化算法研究应用的热点。然而PSO有易早熟和搜索精度不高的缺陷,针对此弊病,在基于仿真的优化(Simulation Based Optimization,SBO)思想体系下,融合人工生命、基于Agent的计算和计算智能,提出面向SBO的PSO计算模型,并将PSO的系统研究和算法改良抽象为一个大规模组合优化问题的求解。最后利用一系列benchmark函数进行了仿真优化实验,取得了较好的结果,从而论证了本思想方法的可行性与可信性。Particle swarm optimization （PS0） is the typical delegate of swarm intelligence which is provided of the characteristics that the parameter number is few and the basic operation is simple, so it has been the hotspot of the optimization domain. At the same time, PSO is provided with the defects that are premature convergence and searching precision. Aiming at the deficiencies, artificial life, agent-based computing and computational intelligence into one within the framework of simulation based optimization （SBO） were integrated and fused. The PSO computational model was proposed using SBO subsequently,＂ moreover, the systematic study and algorithm improvement were abstracted as the solving of the large scale combinatorial optimization problem accordingly. Finally, a series of benchmark functions were tested by the computational model and obtained the high performance, which was supposed to demonstrate the feasibility and creditability of the above methodology.

关 键 词：人工生命 群集智能 粒子群优化算法 组合优化 基于仿真的优化 计算模型 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]